The long-term goals of this project are: 1) to evaluate the feasibility of generating live attenuated virus vaccines of HIV-1 and HIV-2 that are rendered non-pathogenic by mutation of accessory genes, either individually or in combination; 2) to explore the possibility that the accessory gene proteins can be targets for anti-viral therapy. and 3) to generate retroviruses that replicate without integrating into the host genome. As prerequisites to the development of candidate live attenuated virus vaccines and the development of anti-HIV drugs directed against the accessory gene products, we have been engaged on studies to determine the role of these proteins in the life cycle of HIV-1 and HIV-2 in vitro, since a knowledge of how they function is critical to both goals. Our earlier work had demonstrated the critical role of HIV-1 Vif to virus replication in primary T cells (peripheral blood mononuclear cells, PBMC) and in primary monocyte-derived macrophages (MDM). In the case of Nef, we have shown that whether or not Nef has a measurable effect on virus replication depends on the particular virus-host system used. While Nef mutants of several HIV-1 strains all replicate slightly less well than wild type in PBMC and in MDM, there can be either no effect or dramatic reductions in virus replication when Nef mutants of HIV-1 and HIV-2 are assayed in CD4-positive cell lines. As part of our goal to develop attenuated HIV vaccine candidates, we have modified an HIV-1 genome to allow the insertion of different genes into the nef open reading frame. The vif genes of HIV-1 and HIV-2 have been inserted into the both the homologous and heterologous viruses with the aim of determining whether ectopic expression of Vif is functional and whether Vif of HIV-1 can complement Vif mutants of HIV-2 and vice versa. Preliminary work has demonstrated that HIV-2 vif can complement HIV-1 Vif mutants, but that vif from SIVagm cannot. This work will be extended to a study of Nef and Vpr. To explore the feasibility of developing a retrovirus that is replication competent but does not integrate as an obligatory step in its life cycle, we have constructed a murine leukemia virus (MLV) derivative that has an amphotropic env gene as well as mutations in the integrase gene and the terminal repeats; these latter mutations render the virus integration defective. In addition, a replication origin,ori, from the DNA virus SV40 was inserted into the viral genome. The resulting virus, MLVori, was used to infect COS7 cells, which contain the SV40 replication protein T antigen. Virus replication was observed in COS7 cells but not in CV1 cells, which do not contain SV40 T antigen. Although the virus replicated at the initial passage, multiple passages were required before a level of replication was achieved that exceeded that of the parent amphotropic MLV. The determinants of this improved replication capacity will be mapped. This approach is being extended to HIV.